The 2016 U.S. Environmental Protection Agency (EPA) Integrated Risk Information System (IRIS) assessment for ethylene oxide (EO) estimated a 10(-6) increased inhalation cancer risk of 0.1 parts per trillion, based on National Institute of Occupational Safety and Health (NIOSH) epidemiology studies of sterilization facility workers exposed to EO between 1938 and 1986. The worker exposure estimates were based on a NIOSH statistical regression (NSR) model validated with EO levels measured after 1978. Between 1938 and 1978, when EO data was unavailable, the NSR model predicts exposures lowest in 1938 increasing to peak levels in 1978. That increasing EO concentration trend arose, in part, because engineering/industrial-hygiene (E/IH) factors associated with evolving EO-sterilization equipment and operations before 1978 were not properly considered in the NSR model. To test the NSR model trend prediction, a new E/IH-based model was developed using historical data on EO kill concentrations, EO residue levels in sterilized materials, post-wash EO concentrations in a sterilization chamber, and information on facility characteristics and sterilizer operator practices from operators familiar with pre-1978 industry conditions. The E/IH 90th percentile of 8 h time-weighted average EO exposures (C-90) for highly exposed sterilizer operators was calibrated to match 1978 C-90 values from the NSR model. E/IH model C-90 exposures were estimated to decrease over time from levels 16 and were four-fold greater than NSR-estimated exposures for workers during 1938-1954 and 1955-1964. This E/IH modeled trend is opposite to that of NSR model predictions of exposures before 1978, suggesting that EPA's exclusive reliance on the NIOSH cohort to estimate EO cancer risk should be re-examined.
